
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.impute import SimpleImputer
from sklearn.preprocessing import MinMaxScaler
import plotly.express as px
import plotly.graph_objects as go


url ='https://raw.githubusercontent.com/akkuebler/ivda_dataprocessing/main/preppedStudentCovidData.csv'
df = pd.read_csv (url, index_col=0)


df[:5]


print("The dataset contains {} data records and {} features.".format(df.shape[0], df.shape[1]))

The dataset contains 10000 data records and 161 features.


# summary statistics
relevant_questions = ['Q1', 'Q7', 'Q4', 'Q17', 'Q5']
df_relevant = df[relevant_questions]

df_25 = df.filter(like='Q25')
df_26 = df.filter(like='Q26')

df_relevant = pd.concat([df_relevant, df_25, df_26], axis=1)
df_relevant


df_relevant.to_csv('relevant.csv')


df_relevant.describe()


# Which question has the most missing data?
sum_nan = df_relevant.isna().sum()
print(sum_nan.idxmax())

Q17


# Check for duplicates
df_relevant = (df_relevant.reset_index()
        .drop_duplicates(keep='first')
        .set_index('index'))
#df_relevant.drop_duplicates(keep='first', inplace=True, ignore_index=False)
df_relevant


print("After handling duplicates, the dataset now contains {} data records and {} features.".format(df_relevant.shape[0], df_relevant.shape[1]))

After handling duplicates, the dataset now contains 8438 data records and 25 features.


c = list(df_relevant["Q1"])
def counting_as_dict(c):
    countries = dict()
    for country in c:
        if country not in countries:
            countries[country] = 1
        else:
            countries[country] += 1
    return countries
    
    

def plot_pie(df):
    # Created the plot from a data frame, since I was not able to modify the legends of the hoover data otherwise.
    fig = px.pie(df, values='Responses', names='Country', title='Responses from each country')
    fig.update_traces(textposition='inside', textinfo='percent+label')
    fig.show()

country_dict = counting_as_dict(c)

data_list = [(key, value) for key, value in country_dict.items()]
country_df = pd.DataFrame(data_list, columns=['Country', 'Responses'])
plot_pie(country_df)


url2 ='https://raw.githubusercontent.com/akkuebler/ivda_dataprocessing/main/continents.csv'
continents = pd.read_csv (url2, index_col=0)
continents


continents_dict = continents.to_dict()['continent']


# Mapping 
def add_continent(country):
    # The countries of the dataset do not match entirely with the countries in the continent dataset.
    # Example: 'The Netherlands' and 'Netherlands'. That's why unknown continent exist.
    try:
        return continents_dict[country]
    except:
        return "Unknown"

country_df['Continent'] = country_df['Country'].map(lambda x: add_continent(x))

def plot_pie2(df):
    # Created the plot from a data frame, since I was not able to modify the legends of the hoover data otherwise.
    fig = px.pie(df, values='Responses', names='Continent', title='Responses from each continent')
    #fig.update_traces(textposition='inside', textinfo='percent+label')
    fig.show()

#country_dict = counting_as_dict(c)
plot_pie2(country_df)


country_df.to_csv('temp.csv')


country_df.dropna(subset=['Country'], inplace=True)

country_df['Country']=country_df['Country'].str.replace("United States of America", "United States")
country_df['Country']=country_df['Country'].str.replace("The Philippines", "Philippines")
country_df['Country']=country_df['Country'].str.replace("Phillipines", "Philippines")
country_df['Country']=country_df['Country'].str.replace("Democratic Republic of Congo", "DR Congo")
country_df['Country']=country_df['Country'].str.replace("The Netherlands", "Netherlands")
country_df['Country']=country_df['Country'].str.replace("Palestinian State", "Palestine")


# Mapping 
def add_continent(country):
    return continents_dict[country]


country_df['Continent'] = country_df['Country'].map(lambda x: add_continent(x))

def plot_pie2(df):
    # Created the plot from a data frame, since I was not able to modify the legends of the hoover data otherwise.
    fig = px.pie(df, values='Responses', names='Continent', title='Responses from each continent')
    #fig.update_traces(textposition='inside', textinfo='percent+label')
    fig.show()

#country_dict = counting_as_dict(c)
plot_pie2(country_df)


# Visualize 
fig = px.histogram(df_relevant, x="Q7", title="Histogram - age of students", marginal='box', labels={
                     "Q7": "Age",
                 })
fig.show()


fig = px.box(df_relevant, y="Q7", title="Box plot - age of students", labels={
                     "Q7": "Age",
                 })
fig.show()


# Filter data using np.isnan 
Q7_no_nan = df_relevant.dropna(subset=['Q7'])

# Visualize 
fig1 = px.histogram(Q7_no_nan, x="Q7", title="Histogram - age of students", marginal='box', labels={
                     "Q7": "Age",
                 })
fig1.show()

fig2 = px.box(Q7_no_nan, y="Q7", title="Box plot - age of students", labels={
                     "Q7": "Age",
                 })
fig2.show()


print("The dataset before dropping the rows contains {} data records and {} features.".format(df_relevant.shape[0], df_relevant.shape[1]))

# drop rows with missing data
df_relevant.dropna(how='all', inplace=True)

print("The dataset now contains {} data records and {} features.".format(df_relevant.shape[0], df_relevant.shape[1]))

df_relevant

The dataset before dropping the rows contains 8438 data records and 25 features.
The dataset now contains 8381 data records and 25 features.


print("Original Data : \n", df_relevant.Q7.describe(), "\n")


# impute missing values
df_imputed = df_relevant['Q7'].replace(np.nan, df_relevant["Q7"].mean())
print("Imputed Data : \n", df_imputed.describe())

# It is not relevant to describe if we drip NaNs or not since they are ignored, therefore is expected to obtain the same statistics.

# Visualize 
# fig1.add_vline(x=df_relevant["Q7"].mean(), line_dash = 'dash', line_color = 'firebrick', annotation_text="average")
# fig2.add_hline(y=df_relevant["Q7"].mean(), line_dash = 'dash', line_color = 'firebrick', annotation_text="average")
# fig1.show()
# fig2.show()

Original Data : 
 count    6805.000000
mean       23.583688
std         5.680701
min        18.000000
25%        20.000000
50%        22.000000
75%        24.000000
max        70.000000
Name: Q7, dtype: float64 

Imputed Data : 
 count    8381.000000
mean       23.583688
std         5.118729
min        18.000000
25%        21.000000
50%        23.000000
75%        24.000000
max        70.000000
Name: Q7, dtype: float64


# Visualize 
fig3 = px.histogram(df_imputed, x="Q7", title="Histogram plot - age of students with imputed data", marginal='box', nbins=70, labels={
                     "Q7": "Age",
                 })
fig3.show()

fig4 = px.box(df_imputed, y="Q7",title="Box plot - age of students with imputed data", labels={
                     "Q7": "Age",
                 })
fig4.show()


import plotly.subplots as sp

fig = sp.make_subplots(rows=2, cols=1, shared_xaxes=True, subplot_titles=["Histograms - age of students", "Box plots - age of students"])
#fig = go.Figure()
trace1 = go.Histogram(x=df_imputed, nbinsx=70, marker_color="blue", name="Imputed missing fields, histogram")
trace2 = go.Histogram(x=df_relevant["Q7"], nbinsx=70, marker_color="red", name="Ignored missing fields, histogram")
trace3 = go.Box(x=df_imputed, marker_color="blue", name="Imputed missing fields, box plot")
trace4 = go.Box(x=df_relevant["Q7"], marker_color="red", name="Ignored missing fields, box plot")


fig.add_trace(trace1)
fig.add_trace(trace2)
fig.append_trace(trace3, row=2, col=1)
fig.append_trace(trace4, row=2, col=1)

fig.add_vline(x=df_relevant["Q7"].mean(), line_dash = 'dash', line_color = 'firebrick')
fig.add_vline(x=df_imputed.mean()+0.01, line_dash = 'dash', line_color = 'blue')

fig.add_annotation(text="Mean = {}".format(np.round(df_relevant["Q7"].mean(),2)),
                 x=df_relevant["Q7"].mean(),
                 y=500,
                 ax=100,
                 ay=-60,
                 align="left",
                 font=dict(
                     size=14,
                     color="red"
                 )
            )

fig.add_annotation(text="Mean = {}".format(np.round(df_relevant["Q7"].mean()+0.01,2)),
                 x=df_imputed.mean(),
                 y=1200,
                 ax=100,
                 ay=-60,
                 align="left",
                 font=dict(
                     size=14,
                     color="blue"
                 ))

fig.update_layout(title_text="Overlaped plots", xaxis_title="Age", yaxis_title="Unmber of students", barmode="overlay")
fig.update_traces(opacity=0.4)
fig.show()


# Form new category named "Unanswered"
df_enrollmentType = df_relevant[["Q4", "Q7"]].copy()
df_enrollmentType["Q4"] = df_enrollmentType["Q4"].fillna("Unanswered")

# Map floats to categories to create groups (number and str are not compatible)
df_enrollmentType = df_enrollmentType.replace({"Q4":{1.0:"Full-time", 2.0:"Part-time"}})

# Since we are grouping by ages, I am removing the NaN in age column.
df_enrollmentType.dropna(subset=["Q7"], inplace=True)
df_enrollmentType["Q7"] = df_enrollmentType["Q7"].astype(float)


# Visualize 
groups = df_enrollmentType.groupby(["Q4", pd.cut(df_enrollmentType.Q7,[18,28,39,80], include_lowest=True, right=False)]).count()
groups.index.names = ["enrollment-type", "age"]
groups = groups.reset_index()
groups['age'] = groups['age'].astype(str)
groups = groups.replace({"age":{"[18, 28)":"18-27","[28, 39)":"28-38", "[39, 80)":"39+"}})
sums_age = groups[['age', 'Q7']].groupby(by='age').sum()
#print(sums_age.loc['18-27']['Q7'])

percentage_col=[]
for index, row in groups.iterrows():
    percentage_col.append(np.round((row['Q7']/sums_age.loc[row['age']]["Q7"])*100,2))
groups['percentage'] = percentage_col

fig = px.bar(groups, x="age", y="Q7", color="enrollment-type", text_auto=True, hover_data=['percentage'], title="Enrollment type of different age groups", labels={'Q7':'number of students'})
fig.update_traces(textfont_size=14, textposition="outside", cliponaxis=False)
fig.show()

C:\Users\jorge\AppData\Local\Temp\ipykernel_21260\3255625547.py:14: FutureWarning:

The default of observed=False is deprecated and will be changed to True in a future version of pandas. Pass observed=False to retain current behavior or observed=True to adopt the future default and silence this warning.


# Handle missing data
df_3 = df_relevant[['Q4','Q17']].copy()
df_3["Q4"] = df_3["Q4"].fillna("Unanswered")
df_3 = df_3.replace({"Q4":{1.0:"Full-time", 2.0:"Part-time"}})
df_3.dropna(subset=["Q17"], inplace=True)
df_3


# Visualizations
groups_3 = df_3.groupby(by=['Q17','Q4']).size().reset_index()
fig = px.bar(groups_3, x='Q17', y=0, color='Q17', title="Relationship between workload and enrollment type",
             color_continuous_scale=['green', 'yellow','red'],
             hover_data=[0, 'Q17', 'Q4'],
             pattern_shape="Q4", pattern_shape_sequence=[".", "x", "+"],
             labels={'Q17':'Workload', 'Q4':'Enrollment Type', '0': 'Number of Students'}, height=400)
fig.update_layout(coloraxis_colorbar_x=-0.3)
fig.show()

fig2 = px.bar(groups_3, x='Q17', y=0, color='Q17', title="Relationship between workload and enrollment type",
             color_continuous_scale=["#A4161A","#E5383B","#FFFFFF","#F5F3F4","#B1A7A6"],
             hover_data=[0, 'Q17', 'Q4'],
             pattern_shape="Q4", pattern_shape_sequence=[".", "x", "+"],
             labels={'Q17':'Workload', 'Q4':'Enrollment Type', '0': 'Number of Students'}, height=400)
fig2.update_layout(coloraxis_colorbar_x=-0.3)
fig2.show()


df_4 = df_relevant.filter(like='Q25').copy()
#df_4 = pd.concat([df_relevant['Q5'], df_4], axis=1)

print("The dataset contains {} data records and {} features.".format(df_4.shape[0], df_4.shape[1]))

# Handling missing data

# Remove only when all the row is missing, since there are rows that are only missing a column. I don't think it is fair to 
# remove all the row for just one value. Instead I replaced for 0 so it doesn't affect the ESS
df_4.dropna(how='all', inplace=True)
df_4.replace(np.nan, 0, inplace=True)

print("The dataframe after removing rows with NaN value in the specified columns contains {} data records and {} features.".format(df_4.shape[0], df_4.shape[1]))

# Summarize specific columns

emotions={
    'Joyful':'positive',
    'Hopeful':'positive',
    'Proud':'positive',
    'Frustrated':'negative',
    'Angry':'negative',
    'Anxious':'negative',
    'Ashamed':'negative',
    'Relieved':'positive',
    'Hopeless':'negative',
    'Bored':'negative'
}

df_4.columns = emotions.keys()

positive_tags = [k for k, v in emotions.items() if v=='positive']
negative_tags = [k for k, v in emotions.items() if v=='negative']


df_4['positive'] = df_4[positive_tags].sum(axis=1)
df_4['negative'] = df_4[negative_tags].sum(axis=1)
    
# Normalize new columns
df_4['positive'] = ((df_4['positive']-df_4['positive'].min())/(df_4['positive'].max()-df_4['positive'].min()))
df_4['negative'] = ((df_4['negative']-df_4['negative'].min())/(df_4['negative'].max()-df_4['negative'].min()))

# Calculate final emotional satisfaction score
df_4['ESS'] = df_4['positive']-df_4['negative']

# Group by Q5, calculate average 
q5 = df_relevant['Q5']
q5 = q5.replace({1.0:'Bachelor', 2.0:'Masters', 3.0:'PhD'})

df_4 = pd.concat([df_4, q5], join='inner', axis=1)

means = dict(df_4.groupby(by='Q5').mean('ESS').ESS)

# Visualization 
fig = px.box(df_4, x='ESS', y='Q5', color='Q5', 
            labels={
                'Q5':'Studies level',
                'ESS': 'Emotional Satisfaction Score'
            }, title='Emotion Satisfaction in different Study levels')
# Use a better order
fig.update_yaxes(categoryorder='array', categoryarray=['PhD', 'Masters', 'Bachelor'])
fig.add_vline(x=means['Bachelor'], line_dash='dash', line_color='blue')
fig.add_vline(x=means['Masters'], line_dash='dash', line_color='green')
fig.add_vline(x=means['PhD'], line_dash='dash', line_color='red')

fig.add_annotation(text="Mean = {}".format(np.round(means["Bachelor"],2)),
                 x=means["Bachelor"],
                 y=2,
                 ax=100,
                 ay=-60,
                 align="left",
                 font=dict(
                     size=14,
                     color="blue"
                 )
            )

fig.add_annotation(text="Mean = {}".format(np.round(means["Masters"],2)),
                 x=means["Masters"],
                 y=1,
                 ax=100,
                 ay=-60,
                 align="left",
                 font=dict(
                     size=14,
                     color="green"
                 )
            )

fig.add_annotation(text="Mean = {}".format(np.round(means["PhD"],2)),
                 x=means["PhD"],
                 y=0.02,
                 ax=100,
                 ay=-60,
                 align="left",
                 font=dict(
                     size=14,
                     color="red"
                 )
            )

fig.show()
#

The dataset contains 8381 data records and 10 features.
The dataframe after removing rows with NaN value in the specified columns contains 6272 data records and 10 features.


# Select columns of interest
df_5 = df_relevant.filter(like='Q26')

# Filter European countries
df_5 = pd.concat([df_relevant['Q1'], df_5], join='inner', axis=1)
european_countries = {k: v for k, v in continents_dict.items() if v== "Europe"}
df_5['Q1']=df_5['Q1'].str.replace("The Netherlands", "Netherlands")
def is_europe(x):
    if x in european_countries.keys():
        return 'Yes'
    else:
        return np.nan
    
df_5['Europe'] = df_5['Q1'].map(lambda x: is_europe(x))
df_5.dropna(subset=['Europe'], inplace=True)

df_5.drop('Europe', axis=1, inplace=True)

# Filter lowest ESS (Bachelor) using the Data Frame used to calculate the value.
df_5 = pd.concat([df_5, df_4['Q5']], join='inner', axis=1)
df_5['Q5'].replace('Masters', np.nan, inplace=True)
df_5['Q5'].replace('PhD', np.nan, inplace=True)
df_5.dropna(subset=['Q5'], inplace=True)

# Remove missing data
df_5 = df_5.dropna()
# We no longer need Q1 and Q5
df_5.drop(['Q1', 'Q5'], axis=1, inplace=True)
df_5.to_csv('temp.csv')


from sklearn.cluster import KMeans

k_means = KMeans(init="k-means++", n_clusters=5, n_init=10)
results = k_means.fit(df_5.values)
results

KMeans(n_clusters=5, n_init=10)

KMeans(n_clusters=5, n_init=10)


# Since the number of columns is 10 and I can not create a visualization that is understandable
# for every user in more than 3D I am using PCA to 3D because reduce 10 dims to 2 may imply loosing 
# information
from sklearn.decomposition import PCA
pca = PCA(n_components=3)
components = pca.fit_transform(df_5.values)

df_5_3d = df_5.copy()
df_5_3d['Cluster'] = results.labels_
df_5_3d['PCA1'] = components[:,0]
df_5_3d['PCA2'] = components[:,1]
df_5_3d['PCA3'] = components[:,2]

fig_3D = px.scatter_3d(df_5_3d, x='PCA1', y='PCA2', z='PCA3', color='Cluster', opacity=0.8,
                      hover_data=['Q26a','Q26b','Q26c','Q26d','Q26e','Q26f','Q26g','Q26h','Q26i','Q26j'],
                      labels={
                          'Q26a':'Personal physical health',
                          'Q26b': 'Personal mental health',
                          'Q26c': 'Studying issues (lectures, seminars, practical work)',
                          'Q26d': 'Future education',
                          'Q26e': 'Personal finances',
                          'Q26f': 'Family and relationship',
                          'Q26g': 'Professional career in the future',
                          'Q26h': 'COVID19 or similar pandemic crisis in the future',
                          'Q26i': 'Leisure activities',
                          'Q26j': 'Traveling abroad'
                      },
                      size_max=1,title='Clustering 5 Clusters and  PCA (3D Plot)')

fig_3D.show()


# Try 2 D since it is difficult to inderstand and reduce to 4 clusters
from sklearn.cluster import KMeans

k_means = KMeans(init="k-means++", n_clusters=4, n_init=10)
results = k_means.fit(df_5.values)

from sklearn.decomposition import PCA
pca = PCA(n_components=2)
components = pca.fit_transform(df_5.values)

df_5_2d = df_5.copy()
df_5_2d['Cluster'] = results.labels_
df_5_2d['PCA1'] = components[:,0]
df_5_2d['PCA2'] = components[:,1]

fig_2D = px.scatter(df_5_2d, x='PCA1', y='PCA2', color='Cluster', opacity=0.8,
                      hover_data=['Q26a','Q26b','Q26c','Q26d','Q26e','Q26f','Q26g','Q26h','Q26i','Q26j'],
                      labels={
                          'Q26a':'Physical health',
                          'Q26b': 'Mental health',
                          'Q26c': 'Studying issues',
                          'Q26d': 'Future education',
                          'Q26e': 'Finances',
                          'Q26f': 'Family and relationship',
                          'Q26g': 'Future career',
                          'Q26h': 'Pandemic crisis',
                          'Q26i': 'Leisure activities',
                          'Q26j': 'Traveling abroad',
                          'PCA1' : 'x',
                          'PCA2' : 'y'
                      },
                      size_max=1,title='Clustering 5 Clusters and  PCA (2D Plot)')

fig_2D.show()


import plotly
plotly.offline.init_notebook_mode()

	continent
country
Algeria	Africa
Angola	Africa
Benin	Africa
Botswana	Africa
Burkina Faso	Africa
...	...
Paraguay	South America
Peru	South America
Suriname	South America
Uruguay	South America
Venezuela	South America

A3: Data Preprocessing and Knowledge Discovery (20 points)¶

DUE: Sunday, 29 October 2023, 23:59¶

General Task Description¶

Point Distribution:¶

Submission:¶

Task 1 - Handling duplicates & mapping data (3 points)¶

Task 1.1¶

Task 1.2¶

Task 1.3¶

Task 1.4¶

Task 2 - Handling and treating missing values (4.5 points)¶

Task 2.1¶

Task 2.2¶

Task 2.2.1¶

Task 2.2.2¶

Task 2.2.3¶

Task 3 - Grouping data & colour coding (2.5 points)¶

Task 4 - Normalizing data & colour coding (3 points)¶

Task 5 - KDD (7 points)¶

Clear plot in 2D (above is just demostrative)¶

#¶

	Q1	Q3	Q4	Q5	Q6	Q7	Q8	Q9	Q10a	Q10b	...	Q38h_2	Q38i_2	Q38j_2	Q38k_2	Q38l_2	Q38m_2	Q38n_2	Q38o_2	Q38p_2	SOURCE
29349	Turkey	1.0	1.0	1.0	1.0	24.0	1.0	2.0	1.0	1.0	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	tr
23415	Portugal	1.0	1.0	1.0	3.0	NaN	1.0	2.0	4.0	5.0	...	1.0	3.0	5.0	4.0	3.0	5.0	1.0	2.0	2.0	pt
5568	Croatia	1.0	1.0	1.0	2.0	20.0	2.0	2.0	3.0	3.0	...	5.0	1.0	1.0	4.0	2.0	2.0	1.0	1.0	2.0	en
10176	Hungary	1.0	1.0	1.0	3.0	21.0	1.0	2.0	3.0	3.0	...	5.0	1.0	3.0	4.0	4.0	3.0	1.0	3.0	1.0	en
2226	Bangladesh	1.0	1.0	2.0	1.0	NaN	1.0	2.0	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	en

	Q7	Q4	Q17	Q5	Q25a	Q25b	Q25c	Q25d	Q25e	Q25f	...	Q26a	Q26b	Q26c	Q26d	Q26e	Q26f	Q26g	Q26h	Q26i	Q26j
count	8024.000000	9493.000000	6655.000000	9430.000000	6817.000000	6798.000000	6800.000000	6799.000000	6799.000000	6804.000000	...	6719.000000	6706.000000	6704.000000	6692.000000	6697.000000	6702.000000	6686.000000	6684.000000	6689.000000	6671.000000
mean	23.546735	1.115559	3.396093	1.242842	2.853601	3.129303	2.688676	3.340197	2.960141	3.270870	...	2.583867	2.806591	3.175865	3.117603	2.944602	3.024172	3.268771	2.600090	2.729406	2.377455
std	5.677028	0.319712	1.227374	0.523890	1.067821	1.121784	1.155630	1.066495	1.083197	1.180578	...	1.242093	1.316116	1.186078	1.254521	1.351050	1.321371	1.269045	1.299526	1.220846	1.339156
min	18.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	...	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
25%	20.000000	1.000000	3.000000	1.000000	2.000000	2.000000	2.000000	3.000000	2.000000	2.000000	...	2.000000	2.000000	2.000000	2.000000	2.000000	2.000000	2.000000	2.000000	2.000000	1.000000
50%	22.000000	1.000000	4.000000	1.000000	3.000000	3.000000	3.000000	3.000000	3.000000	3.000000	...	2.000000	3.000000	3.000000	3.000000	3.000000	3.000000	3.000000	2.000000	3.000000	2.000000
75%	24.000000	1.000000	4.000000	1.000000	4.000000	4.000000	3.000000	4.000000	4.000000	4.000000	...	3.000000	4.000000	4.000000	4.000000	4.000000	4.000000	4.000000	4.000000	4.000000	3.000000
max	70.000000	2.000000	5.000000	3.000000	5.000000	5.000000	5.000000	5.000000	5.000000	5.000000	...	5.000000	5.000000	5.000000	5.000000	5.000000	5.000000	5.000000	5.000000	5.000000	5.000000

	Q1	Q7	Q4	Q17	Q5	Q25a	Q25b	Q25c	Q25d	Q25e	...	Q26a	Q26b	Q26c	Q26d	Q26e	Q26f	Q26g	Q26h	Q26i	Q26j
index
29349	Turkey	24.0	1.0	NaN	1.0	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
23415	Portugal	NaN	1.0	4.0	1.0	4.0	5.0	5.0	5.0	4.0	...	5.0	5.0	4.0	3.0	5.0	5.0	3.0	4.0	5.0	3.0
5568	Croatia	20.0	1.0	5.0	1.0	3.0	3.0	2.0	4.0	4.0	...	1.0	3.0	5.0	5.0	5.0	3.0	5.0	2.0	4.0	1.0
10176	Hungary	21.0	1.0	1.0	1.0	1.0	3.0	3.0	4.0	1.0	...	1.0	1.0	5.0	5.0	2.0	1.0	1.0	1.0	2.0	3.0
2226	Bangladesh	NaN	1.0	NaN	2.0	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
12150	Indonesia	20.0	1.0	4.0	1.0	3.0	2.0	2.0	5.0	3.0	...	3.0	3.0	5.0	5.0	5.0	2.0	5.0	5.0	5.0	5.0
3969	Chile	35.0	2.0	NaN	1.0	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
29241	Turkey	26.0	1.0	3.0	2.0	3.0	4.0	3.0	1.0	1.0	...	1.0	1.0	1.0	4.0	1.0	1.0	2.0	3.0	3.0	4.0
15925	Malaysia	47.0	1.0	NaN	2.0	3.0	4.0	4.0	3.0	1.0	...	1.0	1.0	2.0	2.0	2.0	2.0	NaN	2.0	2.0	2.0
13276	Italy	23.0	1.0	4.0	1.0	3.0	4.0	3.0	4.0	3.0	...	1.0	1.0	2.0	1.0	1.0	3.0	1.0	1.0	1.0	1.0

	Q4	Q17
index
23415	Full-time	4.0
5568	Full-time	5.0
10176	Full-time	1.0
2589	Full-time	3.0
2765	Full-time	4.0
...	...	...
17410	Full-time	5.0
21430	Part-time	4.0
12150	Full-time	4.0
29241	Full-time	3.0
13276	Full-time	4.0